Endothelium-Derived Hyperpolarizing Factor–Mediated Renal Vasodilatory Response Is Impaired During Acute and Chronic Hyperhomocysteinemia
Background— Endothelial dysfunction is an early event in the development of vascular complications in hyperhomocysteinemia. Endothelial cells release a number of vasodilators, including NO and prostacyclin. Several lines of evidence have indicated the existence of a third vasodilator pathway, mediated by endothelium-derived hyperpolarizing factor (EDHF). EDHF is a major determinant of vascular tone in small resistance vessels. The influence of hyperhomocysteinemia on EDHF is unknown. The present in vivo study evaluates the integrity of the EDHF pathway in the renal microcirculation of rats with acute and chronic hyperhomocysteinemia.
Methods and Results— EDHF-mediated vasodilation was evaluated as the renal blood flow (RBF) response to intrarenal acetylcholine during systemic NO synthase and cyclooxygenase inhibition. Acute hyperhomocysteinemia induced by intravenous homocysteine did not affect EDHF-mediated vasodilation. In contrast, intravenous methionine with subsequent hyperhomocysteinemia impaired the EDHF-mediated RBF response. When the methionine infusion was preceded by adenosine periodate oxidized to prevent the cleavage of S-adenosylhomocysteine to homocysteine and adenosine, a similar impairment of EDHF was observed, but with normal homocysteine levels. Animals with chronic hyperhomocysteinemia induced by a high-methionine, low–B vitamin diet during 8 weeks had a severely depressed EDHF-mediated vasodilation compared with those on a standard diet. Endothelium-independent vasodilation to deta-NONOate and pinacidil was not affected in acute and chronic hyperhomocysteinemia, demonstrating intact vascular smooth muscle reactivity.
Conclusions— EDHF-dependent responses are impaired in the kidney of hyperhomocysteinemic rats. Because EDHF is a major regulator of vascular function in small vessels, these findings have important implications for the development of microangiopathy in hyperhomocysteinemia.
Received September 17, 2003; de novo received December 16, 2003; accepted February 5, 2004.
A large body of evidence has indicated that hyperhomocysteinemia portends an increased risk for atherothrombotic cardiovascular disease. The underlying molecular mechanism remains conjectural, however. An early manifestation of atherosclerosis is endothelial dysfunction. Hyperhomocysteinemia is known to be associated with impaired endothelium-dependent vasodilation in both experimental animals1–3 and humans.4–9 The leading mechanism suggested for the adverse vascular effects of homocysteine on endothelial function involves increased oxidant stress with a depletion of bioactive NO.3
Although NO has generally been considered to be the principal mediator of endothelium-dependent relaxations, evidence is mounting that endothelium-derived hyperpolarizing factor (EDHF) is a major determinant of vascular tone, especially in small resistance vessels.10 These vessels control tissue perfusion and thus may be of larger physiological relevance than conductance arteries. The nature of EDHF is still not entirely elucidated.11 Current evidence suggests that EDHF-mediated responses are initiated by activation of endothelial K+ channels with resultant hyperpolarization of endothelial cells. This endothelial hyperpolarization spreads to the underlying smooth muscle layer through myoendothelial gap junctions, or the efflux of K+ from the endothelial cells elicits hyperpolarization of the adjacent smooth muscle cells. Epoxyeicosatrienoic acids likely have a regulatory role in this pathway.11 The contribution of EDHF to relaxation is dependent on vessel size, being more prominent in smaller arteries than in larger ones.10,11 The majority of the studies on the effect of hyperhomocysteinemia on endothelial function were performed in large-conduit arteries such as the brachial artery, where endothelium-dependent vasodilation is largely NO dependent, and therefore a potential effect on EDHF may have been overlooked. Whether hyperhomocysteinemia interferes with the EDHF pathway is currently unknown.
The aim of the present study was to examine the effect of acute and chronic hyperhomocysteinemia on EDHF-mediated vasodilation in vivo. The contribution of EDHF to endothelium-dependent vasodilation is generally evaluated by probing the response to an endothelium-dependent agonist during combined blockade of NO synthase and cyclooxygenase. The renal microcirculation of the rat was selected for this study. This vascular bed is characterized by a large residual response to acetylcholine during NO synthase and cyclooxygenase inhibition, which is abolished by inhibition of gap junctional communication, indicative of a prominent EDHF pathway.10,12
Renal Blood Flow Measurements
The studies were performed in 59 female Wistar rats with a body weight of approximately 250 g (Iffa Credo, Brussels, Belgium), receiving care in accordance with NIH and national guidelines for animal protection. The rats were anesthetized with thiobutabarbital (Inactin, RBI; 100 mg/kg IP). The trachea was intubated, a jugular vein was cannulated for continuous infusion of isotonic saline (3 mL/h) and administration of drugs, and a carotid artery was cannulated for continuous monitoring of arterial blood pressure. The right renal and suprarenal arteries were exposed via a small abdominal incision. The suprarenal artery was cannulated for intrarenal administration of drugs. A blood flow sensor with an inner diameter of 0.5 to 0.7 mm was placed on the renal artery, allowing continuous renal blood flow (RBF) monitoring (T106 flowmeter, Transonic).10,12
The RBF response to intrarenal acetylcholine (Sigma; 1 to 50 ng in bolus), to the NO donor deta-NONOate (Alexis; 16 to 80 μg in bolus), and to the K+-channel opener pinacidil (Sigma; 25 to 125 μg in bolus) was examined. All experiments were performed in the combined presence of systemic NO synthase and cyclooxygenase blockade: NG-nitro-l-arginine methyl ester HCl (L-NAME) (Sigma; 10 mg/kg bolus followed by 20 mg/kg per hour) and indomethacin (Sigma; 4 mg/kg bolus followed by 8 mg/kg per hour). Before administration of the next dose of acetylcholine, deta-NONOate, and pinacidil, RBF was allowed to return to baseline values. The upper limit of the dose-response curve to acetylcholine, deta-NONOate, and pinacidil was chosen as the highest dose that was devoid of systemic blood pressure effects.
The RBF response to acetylcholine, deta-NONOate, and pinacidil in the presence of L-NAME and indomethacin was examined at baseline and 5, 30, and 60 minutes after infusion of l-homocysteine (40 mmol/kg body wt dissolved in 1 mL saline administered over 5 minutes) (n=6) (Figure 1). l-Homocysteine was prepared from its thiolactone form.13 Briefly, 30.7 mg l-homocysteine thiolactone hydrochloride (Sigma) was dissolved in 0.2 mL of 4N NaOH and incubated for 5 minutes at 37°C. Subsequently, Tris-HCl (pH 8.6) and dithiothreitol were added, and the pH was adjusted to 7 to 8. The final concentrations of homocysteine and dithiothreitol were 100 and 20 mmol/L, respectively. To exclude interference by dithiothreitol, the effect of 20 mmol/L dithiothreitol alone dissolved in saline on the RBF response to acetylcholine, deta-NONOate, and pinacidil in the presence of L-NAME and indomethacin was studied (n=6).
The RBF response to acetylcholine, deta-NONOate, and pinacidil in the presence of L-NAME and indomethacin was examined subsequently at baseline, 30 minutes after the infusion of 1 mL saline, 60 and 120 minutes after the infusion of methionine (Sigma; 0.125 g/kg body wt dissolved in 2 mL saline administered over 10 minutes), and 30 minutes after the infusion of 5-methyltetrahydrofolate (Sigma; 800 μg/kg body wt in 1 mL saline administered over 5 minutes) (n=6) (Figure 1). The protocol was repeated after addition of 20 mmol/L dithiothreitol to the methionine solution (n=5).
The RBF response to acetylcholine, deta-NONOate, and pinacidil in the presence of L-NAME and indomethacin was examined subsequently at baseline, 30 minutes after the infusion of adenosine periodate oxidized (Sigma; 20 μmol/kg body wt dissolved in 1 mL saline administered over 5 minutes), 60 and 120 minutes after the infusion of methionine (0.125 g/kg body wt), and 30 minutes after the infusion of 5-methyltetrahydrofolate (800 μg/kg body wt) (n=6) (Figure 1). Adenosine periodate oxidized is a competitive inhibitor of S-adenosylhomocysteine hydrolase and thus blocks the cleavage of S-adenosylhomocysteine to homocysteine and adenosine.
The RBF response to acetylcholine, deta-NONOate, and pinacidil in the presence of L-NAME and indomethacin was examined subsequently at baseline, 30 minutes after the infusion of 1 mL saline, 60 and 120 minutes after the infusion of methionine (0.125 g/kg body wt), and 30 minutes after the infusion of 1 mL saline (n=6) to provide a time control for the administration of 5-methyltetrahydrofolate (Figure 1).
The animals received a diet enriched in methionine and deficient in folate, vitamin B6, and vitamin B12 (Harlan Teklad TD97345, Harlan Teklad) (n=8); a diet enriched in methionine with high levels of folate, vitamin B6, and vitamin B12 (Harlan Teklad TD98002) (n=8); or standard rodent chow (n=8) during 8 weeks (Table 1). Thereafter, the RBF response to acetylcholine, deta-NONOate, and pinacidil was examined in the presence of L-NAME and indomethacin. The experiments were repeated 15 and 30 minutes after the infusion of 5-methyltetrahydrofolate (800 μg/kg body wt).
Methionine concentrations were determined by a reverse-phase high-performance liquid chromatography technique with a SymmetryShield C18 column (Waters, Ettenleur). Samples were deproteinized by sulfosalicylic acid. Norleucine was added as internal standard. Precolumn derivatization was performed with the use of AccQFluor reagent (Waters). For separation, AccQTag eluens (Waters) was applied with a gradient of acetonitrile/water (60%/40%), starting at 35% and increasing to 100%. Quantification was performed by fluorescence detection.
Total plasma homocysteine concentrations were measured with a high-performance liquid chromatography procedure with reverse-phase separation and fluorescence detection, as described previously.14
Data are presented as mean±SEM. The RBF response to the different agonists is expressed as the area under the curve of the change in RBF (mL/min×min), as detailed previously.10,12 ANOVA and paired and unpaired t tests were used as appropriate. The significance level was set at P<0.05.
Acute Hyperhomocysteinemia by Intravenous l-Homocysteine
The intravenous administration of l-homocysteine resulted in a steep rise of total homocysteine levels, as follows: 12.9±1.0 μmol/L before, 183.4±50.9 μmol/L 5 minutes after, 101.5±28.2 μmol/L 30 minutes after, and 53.3±18.5 μmol/L 60 minutes after bolus administration. Methionine levels did not change, as follows: 46.3±4.7 μmol/L before, 57.0±3.0 μmol/L 5 minutes after, 58.3±3.7 μmol/L 30 minutes after, and 53.7±3.8 μmol/L 60 minutes after bolus administration.
The RBF response to acetylcholine during L-NAME and indomethacin infusion was not different before and at different time points after the administration of l-homocysteine, however (Figure 2A). Similarly, the RBF responses to deta-NONOate and pinacidil were unaffected by the l-homocysteine infusion (data not shown). Dithiothreitol alone did not alter the RBF response to acetylcholine, deta-NONOate, and pinacidil (data not shown).
Acute Hyperhomocysteinemia by Intravenous Methionine
Intravenous administration of methionine resulted in a pronounced rise of the plasma methionine levels and a moderate rise of total homocysteine concentrations (Table 2). When the administration of methionine was preceded by the infusion of adenosine periodate oxidized, a competitive inhibitor of S-adenosylhomocysteine hydrolase, a similar increase in methionine levels was observed, but total homocysteine levels did not rise (Table 2).
The RBF response to acetylcholine during systemic NO synthase and cyclooxygenase blockade was significantly lower 60 and 120 minutes after methionine infusion. Administration of 5-methyltetrahydrofolate resulted in a partial recovery of the RBF response to acetylcholine (Figure 2B). A similar suppression of the RBF response to acetylcholine by methionine and a partial restoration by 5-methyltetrahydrofolate were observed in the animals that were pretreated with adenosine periodate oxidized (Figure 2C). The adenosine periodate oxidized infusion itself had no hemodynamic effects (Figure 2C). Administration of saline instead of 5-methyltetrahydrofolate did not affect the RBF response to acetylcholine (Figure 2D), indicating that the partial recovery after 5-methyltetrahydrofolate administration is not due to the time lag after methionine infusion. Addition of dithiothreitol to methionine resulted in a suppression of the RBF response to acetylcholine similar to that of methionine alone (data not shown). The RBF responses to both deta-NONOate and pinacidil during systemic NO synthase and cyclooxygenase blockade were unaffected by methionine exposure with or without pretreatment with adenosine periodate oxidized (Figure 3).
Animals receiving a high-methionine, low–B vitamin diet during 8 weeks developed a progressive hyperhomocysteinemia (Table 3). Those fed a high-methionine, high–B vitamin diet had lower total homocysteine levels, but they were still significantly elevated compared with the control group (Table 3). Methionine levels were significantly higher in both animal groups receiving an excess methionine diet compared with those on standard rodent chow (Table 3).
The RBF response to acetylcholine during systemic NO synthase and cyclooxygenase blockade was suppressed in the high-methionine, low–B vitamin diet group compared with the standard rodent chow group and those fed a high-methionine, high–B vitamin diet (Figure 4A). The acute administration of 5-methyltetrahydrofolate was unable to restore the impaired EDHF-mediated vasodilation in the high-methionine, low–B vitamin diet group (Figure 4B). The RBF responses to both deta-NONOate and pinacidil were not different in animals fed a high-methionine, low–B vitamin diet or a high-methionine, high–B vitamin diet compared with those fed a standard diet (Figure 5).
In resistance vessels, the EDHF pathway may be at least as important as or even more important than NO in mediating endothelium-dependent vasodilation. In the renal microcirculation, EDHF is known to represent a considerable part of endothelium-dependent responses.10–12 We evaluated EDHF-mediated vasodilation in the kidney as the NO synthase– and cyclooxygenase-independent component of acetylcholine-induced increase in RBF. Incomplete inhibition of NO has been excluded by the abolishment of this response by connexin-mimetic peptides that are known to block EDHF-mediated signal transduction but not NO-mediated vasodilation.10
The salient observation of the present study is that the L-NAME and indomethacin-resistant vasodilation in response to acetylcholine is profoundly impaired in the renal microcirculation of rats with acute and chronic hyperhomocysteinemia. This defect cannot be explained by a nonselective impairment of vascular smooth muscle relaxation because, in both models of hyperhomocysteinemia, vasodilation in response to pinacidil and deta-NONOate was not influenced under the same conditions.
In the methionine-homocysteine cycle, methionine is first transformed to S-adenosylmethionine, which is an essential methyl donor. Transmethylation yields S-adenosylhomocysteine and a methylated acceptor, including DNA and proteins. S-Adenosylhomocysteine is hydrolyzed to homocysteine and adenosine. A methionine load will thus result in a rise of S-adenosylhomocysteine levels and subsequent hyperhomocysteinemia. In the present study, acute intravenous or chronic oral methionine loading resulted in a profound inhibition of the EDHF-mediated renal vasodilation. To address the question of whether the interference with the EDHF pathway was caused by homocysteine itself or by another metabolite, additional experiments were performed. Systemic l-homocysteine infusion was unable to affect EDHF-mediated renal vasodilation, although homocysteine levels rose steeply. Conversely, pretreatment with adenosine periodate oxidized prevented the methionine-induced rise in homocysteine levels but not the inhibition of the EDHF pathway. Adenosine periodate oxidized is a competitive inhibitor of S-adenosylhomocysteine hydrolase and thus blocks the conversion of S-adenosylhomocysteine to homocysteine and adenosine. Under these circumstances, a methionine load will result in elevated S-adenosylhomocysteine but not homocysteine levels. Taken together, these results suggest that homocysteine itself does not cause the endothelial toxicity. Another component of the methionine-homocysteine cycle may be responsible for the observed effects.
Hyperhomocysteinemia can be corrected by folate treatment. Methyltetrahydrofolate, the active form of folate, provides a methyl group in the remethylation of homocysteine to methionine. Folate therapy thus forces the homocysteine-methionine cycle through the remethylation pathway, resulting in an improved ratio of S-adenosylmethionine to S-adenosylhomocysteine. 5-Methyltetrahydrofolate partially corrected the abnormalities in the EDHF pathway during the methionine load, without affecting the rise in homocysteine levels. Acute administration of 5-methyltetrahydrofolate, however, was unable to improve EDHF-mediated vasodilation in chronically hyperhomocysteinemic rats, suggesting a more profound impairment in endothelial function in these animals. In contrast, chronic dietary supplementation of folate, vitamin B6, and vitamin B12 partially prevented the development of hyperhomocysteinemia and the associated endothelial dysfunction induced by the methionine enrichment of the diet. Because restoration of endothelial dysfunction is a surrogate end point for reduction of cardiovascular risk, these data support a role for B vitamins in the prevention and therapy of cardiovascular disease.
Although the effect of homocysteine on the EDHF pathway has not been studied previously, indirect evidence has suggested that hyperhomocysteinemia may interfere with EDHF. Mice heterozygous for a cystathionine β-synthase gene disruption with mild hyperhomocysteinemia are characterized by an attenuated acetylcholine-induced aortic relaxation.3 However, these animals demonstrate a much more pronounced endothelial dysfunction in the mesenteric microcirculation, with even a paradoxical vasoconstriction after methacholine or bradykinine.3 The mesentery is a vascular bed known for his pronounced EDHF activity.15 Although these findings were interpreted as impaired nitric oxide bioactivity, they may be explained by an additional abnormality in the EDHF pathway.
Although the exact pathophysiological role of EDHF requires further characterization, it is known to control microvascular resistance and tissue perfusion.16 The present findings thus have important implications for the development of microvascular disease in hyperhomocysteinemia. Other risk factors for atherothrombotic cardiovascular disease have also been reported to affect the EDHF pathway, including hypertension,17,18 hypercholesterolemia,19 diabetes mellitus,12 and aging.20 Interference with the integrity of the EDHF pathway may thus be a final common pathway through which these risk factors cause microangiopathy and end-organ damage.
In conclusion, EDHF-mediated vasodilation is impaired in the renal microcirculation of acute and chronically hyperhomocysteinemic rats. Rather than homocysteine itself, another component of the methionine-homocysteine cycle may be responsible for the endothelial dysfunction.
This study was supported by the Fund for Scientific Research Flanders and the Fund for Research of Ghent University. Dr Kluijtmans is a postdoctoral fellow of the Netherlands Heart Foundation (1999T023). Dr Blom is an established investigator of the Netherlands Heart Foundation (D97021). The authors thank Tommy Dheuvaert, Julien Dupont, Nele Nica, Mieke Van Landschoot, and Marie-Anne Waterloos for their expert technical assistance.
Woo KS, Chook P, Lolin YI, et al. Hyperhomocyst(e)inemia is a risk factor for arterial endothelial dysfunction in humans. Circulation. 1997; 96: 2542–2544.
Tawakol A, Omland T, Gerhard M, et al. Hyperhomocyst(e)inemia is associated with impaired endothelium-dependent vasodilation in humans. Circulation. 1997; 95: 1119–1121.
Bellamy MF, McDowell IF, Ramsey MW, et al. Hyperhomocysteinemia after an oral methionine load acutely impairs endothelial function in adults. Circulation. 1998; 98: 1848–1852.
Chambers JC, McGregor A, Jean-Marie J, et al. Demonstration of rapid onset vascular endothelial dysfunction after hyperhomocysteinemia: an effect reversible with vitamin C therapy. Circulation. 1999; 99: 1156–1160.
Kanani PM, Sinkey CA, Browning RL, et al. Role of oxidant stress in endothelial dysfunction produced by experimental hyperhomocyst(e)inemia in humans. Circulation. 1999; 100: 1161–1168.
Chao CL, Kuo TL, Lee YT. Effects of methionine-induced hyperhomocysteinemia on endothelium-dependent vasodilation and oxidative status in healthy adults. Circulation. 2000; 101: 485–490.
Van de Voorde J, Vanheel B, Leusen I. Endothelium-dependent relaxation and hyperpolarization in aorta from control and renal hypertensive rats. Circ Res. 1992; 70: 1–8.
Fujii K, Tominaga M, Ohmori S, et al. Decreased endothelium-dependent hyperpolarization to acetylcholine in smooth muscle of the mesenteric artery of spontaneously hypertensive rats. Circ Res. 1992; 70: 660–669.
Eizawa H, Yui Y, Inoue R, et al. Lysophosphatidylcholine inhibits endothelium-dependent hyperpolarization and N omega-nitro-l-arginine/indomethacin-resistant endothelium-dependent relaxation in the porcine coronary artery. Circulation. 1995; 92: 3520–3526.